Jérémie Zimmermann 🎵 💾 ⁉ is a user on mamot.fr. You can follow them or interact with them if you have an account anywhere in the fediverse. If you don't, you can sign up here.
Jérémie Zimmermann 🎵 💾 ⁉ @jz

How much I like the concepts, the the spirit and the code of itself, the absence of full-text search in toots is such a show-stopper for me... :/

Stuff I see pass from the corner of my eye and that I am unable to find again (end up looking for it in f#@$ twitter!), stuff I remember seeing times ago... stuff I want to keep an eye on (and where people not always use hashtags)..

Hope this gets improved sometimes... <3

· Web · 8 · 5

@jz (et en vrai, moi je suis contente que ça soit le cas et qu'un stalker ne puisse pas exhumer mes vieux toots pour me nuire)

@thaega Tout est a priori déja indexé par les moteurs de recherche.. En vrai il suffit d'en utilieser un pour trouver ce que l'on cherche.

Lorsque l'on publie de l'information par définition on la rend *publique*, et espérer qu'elle ne sera pas facilement trouvable comme moyen de protéger qui que ce soit contre quoi que ce soit risque d'engendrer de grosses déconvenues je crois...

Il pourrait y avoir des options instance-wide pour avoir tout ou certains users non-indexés mais même...

@jz @thaega L'indexation peut être désactivée à volonté, l'option est dispo. Et donc, l'idée, c'est d'empêcher le harcèlement (genre chercher "gay" et aller cracher de l'homophobie sur celles et ceux qui sont ouvertement gays). Pour moi, c'est un gros plus justement :) !

@Cyphergoat @thaega on peut le faire tout de même via les moteurs de recherches, ou un tout petit bout d'huile de code... Donc je ne suis pas sûr que réduire la visibilité d'information publiée (donc publique) puisse aider qui que ce soit sur la durée... plutot qu'encourager l'awarness sur les consequences de publication...

àmha les pertes sont réelles, et les bénéfices tout à fait théoriques...

@Cyphergoat @thaega penser empêcher le harcèlement en n'indexant pas de l'information publiée est l'équivalent de "security through obscurity" je crois... une illusion cachant des désastres à venir... :/

@jz @thaega Bah justement, l'indexation sur les moteurs de recherche est désactivable dans Mastodon au choix de l'utilisateur⋅ice :). Après, tu peux toujours faire une PR pour demander une option qui autorise la recherche, ou quelque chose du genre, en prenant en compte que chacun⋅e peut autoriser ou non ces recherches sur son profil, ça devrait pas entrer en conflit avec la logique et ça sera moins "obscurity"-like :).

@Cyphergoat @jz après c'est vrai qu'on peut toujours me retrouver en creux via les profils de mes contacts qui n'ont pas désactivé l'indexation
Mais je pense que ça reste moins pire que Twitter

@jz @Cyphergoat oui et non, si la non indexation n'empêche pas dans l'absolu le stalkage, ça le rend moins aisé, et du coup ça doit arrêter au moins une partie des casse-couilles
Mais c'est vrai que c'est dangereux de partir du principe que ça protège complètement

@thaega @jz Si quelqu’un veut te nuire, il lui faut 10min pour coder un scrapper pour choper tous tes toots hein…

@thaega @jz Ça prend 10min pour trouver quelqu’un pour le coder.

@thaega @jz Pour quelqu’un qui te veut du mal ? Sachant qu’il a déjà Google de base
google.fr/search?source=hp&q=s
Les stalkers aiment se donner beaucoup de mal pour en faire…

@jz @thaega Encore hier, j’suis tombé sur une bande de tarés qui ont écrémé le web pour lever l’identité, l’arbre généalogique, les coordonnées perso, les n° de téléphone, de 182 youtubers allemands.
T’imagines bien que le manuel de curl ne devrait pas trop leur faire peur pour trouver comment crawler tous tes toots.

@jz @thaega Ça me prendrait environ le même temps d’ajouter une petite couche de nokogiri pour parser le html obtenu et l’indexer dans une petite base de données que je pourrais facilement fouiller à la demande…

@jz Search is seen as a double-edged around here, allowing digging into nasty things.
But search of personal post / favorite-ed ones might get in in the near future !

@Technowix By definition what is published is *public* (and therefore indexed by search engines and other parties..).. thinking that hiding public information will help protect people is a bit irresponsible I think.

Only helping there being aware of the consequences of what they publish can truly, durably help, I think.

One could think of instances/profiles with disabled indexing (like through robots.txt) but I think default should be open to maximize access/sharing/visibility...

@jz @Technowix already exists in fact

Settings > Preferences > Opt-out of search engine indexing

@Technowix @jz "One could think of instances/profiles with disabled indexing (like through robots.txt)"

@TheKinrar @jz Yeah ? Actually, the idea was more to "opt out" those who have selected this option of the "local" search engine ^^

@Technowix @TheKinrar I wouldn't mind the "do not index" to be ticked by default, with a clear explanation of what it means to be read (and even a confirmation?) before de-activating.

But this would mean that we actually *could* search in the toots that are free-to-index, which is the missing feature for me here...

@jz @Technowix @TheKinrar why not just patching your instance to add fts ? It's available on mastodon.host GitHub if you want ...

@Technowix @gled @jz @TheKinrar I can see that, but at the same time if you wait for the main 'school of thought' to get the features your users want to increase their engagement, you may wait forever ;)

@Technowix @gled @jz @TheKinrar regarding the search feature ? I've had users asking about it since m.h started... A few also left because of that.. let's see how long it will take to be implemented upstream ;)

@jz yeah, think the same, but believe me some use it in a very personal way... :v

But search of "public" tagged post should be nice, but not planned right now for "protection".
That might evolve of course, but right now the best things is "oh, I liked, paf, favorite"

@jz yeah, same for Diaspora, even for posts I wrote. I end up bookmarking my own posts :D

@jz Yes, this makes Mastodon unsuitable for following current events, or technologies, for instance.

The fact that you cannot even search your own toots also make it unsuitable for PKM (Personal Knowledge Management). Mastodon is more like IRC than like Twitter.

@jz @bortzmeyer Hey, I often find myself grepping through IRC logs to look for things I know I've seen mentioned months ago. So in that regard, even IRC is more useful than Mastodon.

@galaxis @jz May be using the API to keep a log of all your toots? (I don't know if the API allows to retrieve everything in your timeline.)

@bortzmeyer @tom79 includes a backup toot functionnality (beta) in its last release. So it seems possible :) ( 💤 )
@galaxis @jz

@jz @bortzmeyer All your own posts are retrievable, see mastodon-backup by @kensanata at github.com/kensanata/mastodon-

There's also at least one instance that allows full-text search on the Postgres DB through a minor extension, but I don't currently remember which one that was...

@galaxis @jz @bortzmeyer @kensanata at some point groups are planned as a feature which would make this a lot better. I know what you mean as I wanted to find one of my own toots of illustrating some self modifying code on Z80 but didn't hashtag it correctly so its down the rabbithole now..

@bortzmeyer @jz I think this is more of an anti-harassment feature. Avoids people easily searching for topic X and then targeting the users who mention it. You can still use other tools to search through public posts.

@bob @bortzmeyer That's precisely my point: If it is meant to "protect" then it could be tragically counter-productive (as anyone can actually search with a little bit of effort) like "security through obscurity"... when a protective default, *teaching about the implications of publications, etc.* but with the ability for anyone to *choose* to publish content searchable for everyone may turn Mastodon into a tool useful for more than chitchat (research, watch and the such)...

@jz @bortzmeyer Yes, having search as an admin option would make sense, although frustrating research of the fediverse might be a feature. I like the idea of these spaces being contextual and therefore difficult to totalize - similar to Scuttlebutt.

On the protection point I assume that Twitter trolls operate mainly in the most knuckle-dragging manner with no technical sophistication, just by searching for topics or users and then dog-piling them.
@jz @bortzmeyer I think it would be nice if the primary way of searching the fediverse was via having real social relations, rather than being able to bypass social relations with technical methods. It could be argued that the technical bypass of the social is one of the main problems with the centralized social networks.

@bortzmeyer #mastodon Nothing is perfect, but at least you can search for your own toots.

@Italcaus Doesn't work. (Zero results for "@bortzmeyer bgp" and I wrote many toots about BGP)

@bortzmeyer All your toots found but not sorted using a specific topic

@Italcaus This is what I said: you cannot search your own toots.

I guess you could query the database for it as an admin or build a tool around this, but that's not really convenient.
reading the rest of the thread: the UI of Mastodon really describes things in a way they are not. For example if you opt-out elsewhere, there is no X-No-Archive we agreed on in federation, so it'll be public and indexed elsewhere.

@jz

Yep.
One thing that particularly peeves me is when people link to an article - usually I open the article right away but can't read it right now.
And when I find time to read it, often I want to comment one or another particular aspect, but the original toot is drowned under a zillion toots and there's no practical way to find it....